Petasense: This tiny startup offers a big push to Modi's 'Make in India'
Petasense is one of the 10 most disruptive machine-learning startups in the world, tracking a machine's health just as a Fitbit tracks a user's activity.
Launch date: January 2014
Founding team: Abhinav Khushraj and Arun Santhebennur
Location: San Jose, California (HQ) & Bangalore
Funding announced in 2017: $1.8 million
Overall funding raised since starting up: $1.8 million
Strength of the company: 15
Industry it operates in: Industrial IoT
A start-up run by just 15 people can power lumbering Indian factories to higher production. Petasense brings machine-learning and Internet of things (IoT) to industrial-age smokestacks, promising to liberate shop floors from crippling inefficiency with smart technology.
When it comes to health, you have Fitbit that monitors your activities and helps you stay fit. Other devices like Oximeter, a blood glucose monitor, help you track your crucial health parameters. Now, a startup wants to predict the health of a machine and provide early warnings about potential problems.
"Everything from the food we eat to the electricity that lights our homes is powered by factories," says Abhinav Khushraj, co-founder of Petasense, an industrial IoT startup. "These factories have millions of sophisticated machines that perform a menagerie of functions, none—or very few—of which are connected to the Internet," he adds.
According to Khushraj, plants that are woven around IoT are more productive than their non-connected counterparts. This is how he breaks it down:
"The general upkeep of machines—be it motors, pumps or compressors—is critical when running a factory. Since little attention is given to maintaining the health of these machines, it can lead to unplanned downtime, which could translate to millions in lost production," he says. "We want to erase these unplanned shutdowns and reduce costs associated with preventive maintenance. Using our software solution, customers will be able to monitor the health of their equipment without the need for manual analysis," he adds.
In short, Petasense vows to make machines smarter by driving inefficiencies out of the system.
Fitbit for industrial machines
Just as a Fitbit tracks a user's activity, Petasense's integrated hardware and software system monitors and predicts the health of critical industrial assets. This system consists of wireless vibration sensors, cloud software and machine-learning analytics.
Interestingly, investors of the fitness wearable have backed Petasense as well.
"We help customers understand when an equipment is in need of maintenance by listening to its vibrations," says Arun Santhebennur, another co-founder of Petasense. "Machines typically vibrate at a certain frequency, but this alters with wear and tear. Equipped with Petasense's wireless sensor (called the Vibration Mote), these machines will be able to feed data about the vibrations wirelessly to the Cloud, analyse it and present it back to the customer," he adds.
The Vibration Mote monitors the health of rotating machinery such as motors, pumps, fans and compressors. This battery-powered device sends data through WiFi to the Cloud, which is stored in a secure and scalable platform hosted on Microsoft Azure.
Thus, Petasense builds IoT-based sensors in Cloud software which studies vibrations using sensors—but that is not all. The core value of Petasense Cloud comes from machine-learning analytics, which provide meaningful insights about machine conditions based on the data collected.
"This gives the customer more control over the maintenance process," Khushraj says. This means that Petasense will help predict when the machine is going to break down well before it actually does, allowing the customer to schedule maintenance work accordingly. Moreover, its web and mobile applications make it easy to track the health of machines from anywhere, anytime, thus simplifying the industrial IoT.
According to the duo, they are well-placed to deliver a seamless user experience since the startup is building its own hardware (Vibration Mote) as well as software (machine learning).
More importantly, Petasense's integrated software-cum-hardware solution comes at an affordable price, starting at just $399. Clients also get the opportunity to run a 2-to-3 month pilot program where they can evaluate the solution for themselves.
Having emerged from stealth mode recently, the three-year-old startup's growing portfolio of clients includes JLL, C&W Services, Silicon Valley Power and Stanford University. It has also been selected by Google as well as leading VC firms, Data Collective and Emergence Capital, as one of the 10 most disruptive machine learning startups, having contended against 350 other companies.
In addition to being led by MIT and IIT graduates who have 25 years of combined experience in building enterprise companies, Petasense also boasts of having a team comprising engineers, marketers and customer-service personnel from companies like Cisco, Nvidia, Nokia, IBM, Chevron and Deloitte Consulting.
"We are seeing great traction. In fact, we have already logged more than 18 billion wireless vibration readings from a million sensor measurements," Khushraj says.
Petasense also raised a $1.8 million seed round led by True Ventures, with Felicis Ventures and several angel investors also participating.
"The work we do creates a positive impact on industries that bring us the essentials of life—from food & beverages to oil & gas," Santhebennur says. "We draw upon different branches of engineering and sciences—be it mechanical engineering, electronics, computer science, mathematics or physics—to offer a solution that seeks to safeguard the health of high-value industrial machinery."
Although Petasense today employs advanced machine-learning algorithms to monitor and predict the health of critical industrial gear, the startup did not arrive at this without bumps and troughs along the way.
"The one challenge that we faced when starting up, and which we incidentally still do, is the FUD—Fear, Uncertainty and Doubt—factor," Khushraj says. "The industrial sector, which has been doing maintenance the traditional way for years, is still conservative and is reluctant to take the leap to disruptive modern ways," he adds.
The innovators, however, are seeing huge tangible benefits and are actively recruiting on the engineering and sales side for both Bangalore and San Jose offices.
"We feel there needs to be more awareness and education around the tremendous benefits of IoT and predictive maintenance," says Santhebennur. "Manufacturing and Industry 4.0 is very strategic for a country like India. With the 'Make in India' campaign, there is an increased need for the manufacturing conglomerates in India to aggressively adopt digital technologies," he adds.
According to the co-founders, it is no longer enough to compete in the global manufacturing market with labour arbitrage alone. They firmly believe that a product like Petasense can immensely help companies like Reliance, ONGC, Tata Steel, Tata Power, Hindalco, NTPC and others to compete with global players on operational efficiencies.
"We are in rapid growth mode at this time and the future seems promising," Khushraj says. "The market for rotating equipment alone is estimated at $10 billion and the broader predictive maintenance market is very large, given that industries have various kinds of assets and machinery. Going forward, customers would be able to monitor all kinds of parameters, not just vibration, and for all kinds of machines, not just rotating ones. A real-time multi-parametric view of the industrial assets will help us achieve this,” he adds.